Toxicologic Pathology, 32:600–612, 2004 CopyrightC by the Society of Toxicologic Pathology ISSN: 0192-6233 print / 1533-1601 online
DOI: 10.1080/01926230490515201
Qualitative and Quantitative Histomorphologic Assessment of Fathead Minnow Pimephales promelas Gonads as an Endpoint for Evaluating
Endocrine-Active Compounds: A Pilot Methodology Study
JEFFREYC. WOLF,1DANIELR. DIETRICH,2URSFRIEDERICH,3JOHNCAUNTER,4ANDANDREWR. BROWN4
1EPL, Inc., 22866 Shaw Road, Sterling, VA 20166, USA
2Environmental Toxicology, University of Konstanz, Germany
3Dow Europe Gmbh, Switzerland, and
4Brixham Env. Laboratory, AstraZeneca UK Ltd., Brixham, U.K.
ABSTRACT
Although histopathology is routinely employed as a tool for the detection and assessment of xenobiotic-mediated effects in mammals, it is less frequently applied to fish. In part, this is due to a lack of method standardization regarding study design, tissue preservation, tissue sectioning, histopathological evaluation, reporting, and statistical analysis. The objectives of the present study were: (1) to test and refine a method for the microsurgical excision of fathead minnow (FHM)Pimephales promelasgonads for the purpose of histopathologic examination; (2) to determine the optimal combination of fixation and embedding procedures for the histopathologic and morphometric analysis of FHM gonads following exposure to a known estrogenic compound, 17β-estradiol (E2); and (3) to provide a method for the categorization and quantification of cell types in FHM gonads by manually counting cells in digitized images using image analysis software. The light microscopic evaluation of individual gametogenic cells was greatly facilitated by specimen preparation techniques that included the excision of gonads via microdissection and by optimized fixation and embedding procedures.
Keywords. Endocrine-active; reproduction; fathead minnow;Pimephales promelas; testis; ova; morphometry; histopathology.
INTRODUCTION
Fish are suitable models for evaluating endocrine-active compounds in the water column. Fish are exposed to wa- terborne chemicals simultaneously via the respiratory, diges- tive, and dermal routes, and the short reproductive cycle of small fish species ensures that chemical effects on reproduc- tion are swiftly apparent. Perhaps most importantly, it has been demonstrated that various fishes undergo physiological and morphological changes in response to putative endocrine disrupters (Christiansen et al., 1998; Miles-Richardson et al., 1999a; Ankley et al. 2001; Van den Belt et al., 2002; van der Ven and Wester, 2002; Kinnberg and Toft, 2003; van der Ven et al., 2003).
A variety of endpoints have been utilized to investigate the effects of hormones and hormone mimics on the fish repro- ductive system. Examples include changes in plasma levels of vitellogenin and sex steroids such as 17β-estradiol and 11-ketotestosterone, alterations in the external morphology of sexually dimorphic fishes, variations in gonado-somatic indices (ratios of gonad weight to body weight), and re- productive fecundity. Although histopathology is routinely employed as a tool for the detection and assessment of xenobiotic-mediated effects in mammals, it is less frequently
1Abbreviations: FHM, fathead minnow; E2, 17β-estradiol; NBF, neutral buffered formalin; GIP, grid intersection points; VC, vacuolated cell; and ABC, apoptotic body cell.
Address correspondence to: Dr. Jeffrey C. Wolf, Experimental Pathol- ogy Laboratories, Inc., 22866 Shaw Road, Sterling, Virginia 20166, USA;
e-mail: jwolfepl@aol.com
applied to fish. In part, this is due to a lack of method stan- dardization regarding study design, tissue preservation, tissue sectioning, histopathological evaluation, reporting, and sta- tistical analysis, primarily related to a lack of regulatory re- quirements mandating the examination of fish tissues. Con- sequently, reliable and readily comparable results have not been achieved consistently.
The histopathological assessment of fish reproductive or- gans can be divided into 2 separate components: the evalu- ation of gonads for abnormal findings (examples of which include such varied observations as necrosis, Sertoli cell hypertrophy/hyperplasia, and the occurrence of testicular oocytes), and gonad staging. Gonad staging involves the as- sessment of germinal cell type proportions in order to identify potential effects of exogenous or endogenous chemicals on gametogenesis. Gonad staging can be performed quantita- tively or semiquantitatively; the latter is more commonly re- ported, and different semiquantitative staging schemes have been described (Miles-Richardson et al., 1999a; Nichols et al., 2001; Jensen et al., 2001; Ankley et al., 2002; Van den Belt et al., 2002; U.S. EPA, 2002).
Because semiquantitative systems primarily assess the vi- sual density of gametogenic precursors compared to mature gametocytes, a shortcoming inherent to such systems is a lim- itation in the types of data that may be obtained. As an exam- ple, the loss or overabundance of a specific gametogenic stage may not be apparent using some semiquantitative systems.
Such shortcomings may be mitigated by using a quantitative staging approach in which gonadal cell types are individually identified and counted. Furthermore, a distinct advantage of 600
Konstanzer Online-Publikations-System (KOPS) URL: http://www.ub.uni-konstanz.de/kops/volltexte/2008/4988/
URN: http://nbn-resolving.de/urn:nbn:de:bsz:352-opus-49880
quantitative histomorphometry is the ability to apply statisti- cal methods to the data in order to discern subtle treatment- related effects that might otherwise be difficult to appreciate.
There are several reports in which quantitative schemes were used for the staging of fish gonads (Smith, 1978; Depart- ment of the Interior, 2000; Sohoni et al., 2001; van der Ven et al., 2003). In contrast to semiquantitative staging, a key requirement for quantitative staging is that each counted cell must be assigned to one of several well-defined categories according to predetermined histomorphologic criteria. Con- sequently, in order to adequately visualize and differentiate the different gonadal cell types, the tissue collection proce- dures and the preparation of the histologic sections need to be optimized—enhanced cellular detail is especially impor- tant for distinguishing the relatively small germinal cells of the fish testis. Therefore, the objectives of the present study were: (1) to test and refine a method for the microdissec- tion of fathead minnow (FHM)1 gonads for the purpose of histopathologic examination; (2) to determine the optimal combination of fixation and embedding procedures for the histopathologic and morphometric analysis of fathead min- now gonads following exposure to a known estrogenic com- pound, 17β-estradiol (E2); and (3) to provide a method for the categorization and quantification of cell types in FHM gonads by manually counting cells in digitized images using image analysis software.
MATERIALS ANDMETHODS
Study Design and Exposure of Fish to the Test Substance The study design is summarized in Table 1. Briefly, 9 male and 9 female fathead minnows,Pimephales promelas,were continuously exposed to 10 nM 17β-estradiol (2780 ng/l) during the 10-day study. Because documentation of 17β- estradiol effects per se was not a primary aim of this pilot project, negative (diluent water) control fish were not incor- porated into the design of the present study, and the overall number of fish was kept to a minimum. All FHM were ap- proximately 160 days posthatch at the start of the study, and the mean length/weight measurements at terminal sacrifice were 64 mm/5.73 g for males, and 47 mm/1.99 g for fe- males. The test substance, 17β-estradiol, was obtained from Sigma Aldrich Chemical Company (product number E8875, lot number 070K1206). The certificate of analysis accom- panying the test substance confirmed a purity of 100%. A single stock solution of 5.56 mg L−1was prepared in ethanol (90% v/v) and was stirred continuously during the test. This stock solution was delivered to a mixing chamber at a nom- inal flow rate of 0.125 ml min−1, along with dilution water, at a nominal flow rate of 250 ml min−1, giving a nominal
TABLE1.—Distribution of fish numbers, fixation methods, and embedding media.
17β-estradiol exposure concentration
Numbers of fish
Method of fixation Embedding medium Males Females
Formalina Paraffin wax 2,780 ng L−1 3 3
Bouin’s solution Paraffin wax 2,780 ng L−1 3 3
Glutaraldehydeb Glycol methacrylate 2,780 ng L−1 3 3
a10% neutral buffered formalin.
b5% glutaraldehyde, unbuffered.
dilution, immediately before delivery into the test vessel, of 2000:1. This dilution gave the required nominal test con- centration of 2780 ng L−1, along with an ethanol (90% v/v) concentration of 500µl L−1. The dilution water was dechlo- rinated tap water supplied from a 100 m3 reservoir with an average retention time of 24 hours. It was passed through acti- vated carbon, coarsely filtered to remove particulate material, dechlorinated with sodium thiosulphate, and salts added, as required, to maintain minimum hardness levels. The treated water was held in a secondary reservoir with a capacity of 36 m3 (Christiansen, Korsgaard, and Jespersen (1998) and an average retention time of 8 hours. The water was passed through an ultraviolet sterilizer to a second set of filters (25 and 10µm mesh size) and then to a third storage tank with a capacity of 13.5 m3(Christiansen, Korsgaard, and Jespersen (1998). The treated water was delivered via a ring circuit to a temperature-controlled header tank in the test laboratory set to a nominal temperature of 25±1◦C and finally passed through a 5µm filter before use. The volume of test solution in which fish were exposed was 57 L, giving a maximum loading of 6.0 g of fish per liter of solution, and the total flow was equivalent to a minimum of 1 L per g of fish per day.
Analysis of 17β-estradiol concentrations was performed us- ing an estradiol enzyme immunoassay kit (Cat. No. 582251, Cayman Chemical Company, Ann Arbor, Michigan, USA).
Calculated from the day 0 to day 10 results, the mean of the measured 17β-estradiol exposure concentration was 90% of the nominal concentration (range, 79–115% of nominal). Ex- posures were conducted at a temperature of 25◦C±1◦C, and the photoperiod was 16 hours of light followed by 8 hours of darkness, with a 20-minute dawn-and-dusk transition period, respectively. The exposure tanks were aerated to maintain sufficient oxygen levels. Test conditions monitored during the study included dissolved oxygen (mg L−1), pH, temper- ature (◦C), dilution water flow (ml L−1), toxicant flow (ml L−1), and dilution ratio. The dilution water (dechlorinated and filtered to 5 µm) was also analyzed for hardness and conductivity once weekly.
Euthanasia and Necropsy
At the end of the test period, each of the 18 fish was in- dividually captured, euthanized by an overdose of tricaine methanesulfonate (MS222, 3-aminobenzoic acid ethyl ester methanesulfonate 500 mg/l; buffered to pH 7–8 using 1M NaOH), followed by cervical severance, and then immedi- ately necropsied. Individual fish were selected for euthanasia and necropsy at random, alternating male with female fish.
Each fish was necropsied by placing it in right lateral re- cumbency on the stage of a dissecting microscope. Fine dis- secting instruments were used to remove the left body wall and to excise the gonads by severing the spermatic ducts or oviducts and mesenteric attachments. All of the gonads were dissected in a caudal to cranial direction. Macroscopic exter- nal or internal lesions were noted for individual fish. The left and right gonads from each fish were placed into separate compartments of a plastic tissue cassette and assigned to- gether to 1 of 3 fixation solutions, in rotating fashion, accord- ing to the order in which the fish were necropsied. The 3 fix- ation solutions used in this study were: 10% neutral buffered formalin (NBF); Bouin’s solution (71% saturated picric acid,
24% formaldehyde, 5% glacial acetic acid; Sigma-Aldrich, St. Louis, Missouri, USA); and 5% glutaraldehyde in deion- ized water (Table 1). At 24 hours following tissue collec- tion, cassettes placed in Bouin’s solution were rinsed in 70%
ethanol and then transferred to containers of 70% ethanol.
Tissue Processing, Embedding, and Microtoming
Left and right gonads from each of the 18 fish were pro- cessed routinely according to standard histologic methods, and then embedded as follows: gonads fixed in formalin or Bouin’s solution were embedded in paraffin (Paraplast, Tyco Healthcare Group, Mannfield, Massachusetts, USA);
whereas, gonads fixed in glutaraldehyde were embedded in glycol methacrylate (GMA, Polysciences, Warrington, Pennsylvania, USA) (Table 1). The gonads in block form were trimmed to their largest longitudinal sectional area, and then sections (each section approximately 3–5µm thickness for paraffin-embedded tissues or 1–2µm thickness for GMA) from each of the 18 fish were microtomed from each gonad, placed on individual glass slides, and stained with hema- toxylin and eosin (H&E).
Qualitative Histomorphologic Assessment
The suitability of the various fixation and embedding com- binations for the identification of individual gonadal cell types was evaluated subjectively via light microscopy. In ad- dition, all gonad sections were examined for morphologic abnormalities and potential treatment-related changes. The severity of any inflammatory, degenerative, or proliferative changes was graded as minimal, mild, moderate, moderately severe, or severe. Macroscopic observations were correlated with microscopic findings whenever possible.
Quantitative Histomorphologic Assessment
This assessment involved the identification of gonad cell types and the counting (manual tagging) of individual go- nad cells in digitized photomicrographs. Equipment used for the acquisition of images included an Olympus BX51 mi- croscope, a SPOT Insight Color digital camera (Diagnostic Instruments, Inc., Sterling Heights, Michigan, USA) plus ver- sion 3.2 of the accompanying SPOT software, and an IBM- compatible Dell Dimension personal computer running the Microsoft Windows 2000 Professional operating system.
Optimal results for the testes were obtained from sections that were preserved in 5% glutaraldehyde and embedded in GMA (see Results section); therefore, manual tagging of testes cell types was restricted to digital photomicrographs of these sections. Four digital images (0.22 mm×0.29 mm subject area) were obtained from each of the left and right testis sections (i.e., 8 images total per fish) using the 40×mi- croscope objective. To ensure the acquisition of nonoverlap- ping areas, an ultrafine permanent marker was used to create a grid (each grid square was 2×2 mm) on the underside of each glass slide (the overall size of the grid was created large enough to fully encompass each testis section). Areas for pho- tography were selected from the centers of grid squares that were chosen arbitrarily. For each digital image, ImagePro- Plus (IPP—Media Cybernetics, Silver Spring, Maryland, USA, version 4.1) was used by a pathologist to manually tag individual testis cells as to cell type. Tagging was performed by a single pathologist to avoid inconsistencies. The contrast
of each image was globally enhanced using IPP’s “Best Fit”
equalization algorithm. A virtual grid consisting of 400 (20× 20) individual intersection points was applied to each image as an overlay. Grid intersection points (GIP) were separated from one another by 0.015 mm horizontally and 0.011 mm vertically.
Each GIP was tagged with 1 of 6 different-colored dots to represent one of the following cell categories (for criteria see the Results section): spermatozoon, spermatid, spermatocyte, spermatogonium, vacuolated cell (VC), or apoptotic body cell (ABC). All GIP were tagged, with the following exceptions:
GIP that covered empty space in extracellular areas, GIP that covered interstitial structures (collagen, blood vessels), or GIP that covered nonspermatogenic cells or cells that could not be identified (e.g., due to tangential sectioning of the cells or tissue artifact). A minimum of 2,000 spermatogenic cells total were counted per male fish. A grid was used to count cells in the testis because it provided a nonbiased method for subsampling a tissue that contains a large number of very small cells. No attempt was made to extrapolate these counts to any type of volumetric (i.e., stereological) measurement.
Optimal results for the ovaries were obtained from sec- tions that were preserved in Bouin’s solution and embedded in paraffin (see Results section); therefore, the manual tag- ging of ovarian cell types was restricted to images of sections made with this procedural combination. Two digital images (2.2 mm×2.9 mm subject area) were obtained from each of the left and right ovarian sections (i.e., 4 images total per fish) using the 4×microscope objective. To ensure the acquisition of nonoverlapping areas, an ultrafine permanent marker was used to draw a line on the underside of the glass slide that approximately bisected the ovary section perpendicular to its long axis. The 2 tissue areas for photography, obtained from each half of the ovary, were selected arbitrarily. For each digital image, IPP was used by a pathologist to manually tag individual ovarian cells as to cell type. The contrast of each image was globally enhanced using IPP’s “Best Fit” equal- ization algorithm. Via the computer mouse, the pathologist then sequentially tagged each ovarian follicle with 1 of 6 dif- ferent solid-colored squares that each represented 1 of the 6 following follicle cell types: perinucleolar, cortical alve- olar, early vitellogenic, late vitellogenic, mature/spawning, and atretic. A minimum of 100 follicles total were counted per female fish. Counts per image of the different cell types were exported to a Microsoft Excel worksheet and percent- ages were calculated for each cell type per gonad (testis or ovary, left and right) and per fish. To avoid having to make geometric assumptions concerning the shapes of cells or or- gans, the results were maintained as proportions (e.g., number of perinucleolar oocytes per total oocytes counted), without any attempt to convert the counts to stereological or volu- metric estimates (e.g., number of perinucleolar oocytes per unit volume of ovary). Because the distribution of oocytes in cyprinid fishes such as fathead minnows is essentially ran- dom, a single section through the gonads is considered to be representative. This is In contrast to the situation in atherini- form fishes (medaka, for example) that have peripheral to central and cranial to caudal gradients of cell development in their gonads; an alternative approach to assessing cell popu- lations would be necessary to avoid a sampling bias for that species.
TABLE2.—Histopathologic correlates to macroscopic findings in the gonads of fathead minnows exposed to 17β-estradiol.
Macroscopic finding Histopathologic correlate
Testis Attenuation No specific correlate
Ovary Attenuation, gelatinous texture Follicular atresia
RESULTS
Necropsy and Macroscopic Findings
As recognized in previous method development exercises, the use of a dissecting microscope was considered to be at least highly advantageous, if not essential, for the removal of FHM gonads to minimize trauma to these delicate tissues.
Additionally, removal of the swimbladder prior to excision of the gonads facilitated the microdissection procedure, and it appeared that trauma was minimized by dissecting the gonads in a caudal-to-cranial direction, while applying very gentle traction to the spermatic ducts or oviducts.
Macroscopic findings (Table 2) included observations per- taining to relative size differences among the testes and ovaries, respectively. For the males, testis size differences could not be attributed to any specific histopathologic diag- noses. For the females, ovarian size was generally dependent upon the degree to which large follicles were present in each ovary. Several of the excised ovaries had a gelatinous appear- ance and texture which were attributed to varying levels of follicular atresia as determined subsequently by microscopic examination.
Qualitative Histomorphologic Assessment
The combination of glutaraldehyde fixation and GMA em- bedding was judged to be superior to the combinations of Bouin’s or formaldehyde and paraffin wax for identifying in- dividual gonad cells in the testis (Figure 1A, 1C, and 1E).
Cytoplasmic and nuclear details of the spermatogonia, VC, and ABC were enhanced in the thin (approximately 1–2µm) GMA-embedded sections, and the cytoplasmic borders of individual cells could often be recognized. In contrast, fix- ation of the testes in Bouin’s solution produced cells with less nuclear and far less cytoplasmic detail, and individual cell borders were usually indistinct due to the greater thick- ness of the paraffin sections (approximately 3–5µm). Over- all, testis cell types could be adequately (but not optimally) distinguished in Bouin’s/paraffin sections. Compared to the other fixatives, 10% NBF was considered to be inadequate for the purpose of cell counting in the testis; consistent dif- ferentiation of individual cell types was virtually impossible due to the artifactual shrinkage and condensation of the sper- matogenic precursor nuclei.
The combination of Bouin’s solution fixation and paraffin embedding was considered to be the most satisfactory of the 3 preparation methods for the ovary (Figure 1B, 1D, and 1F).
The cytological detail of follicles and supportive cells was actually superior in the glutaraldehyde/GMA sections when compared to Bouin’s fixed sections; however, because ovar- ian follicles are vastly larger than spermatogenic cells, this increased level of detail did not facilitate the differential iden- tification of follicle types. In addition, the Bouin’s/paraffin sections provided comparatively better color contrast among the different follicle types and a more classic morphologic ap-
TABLE3.—Qualitative histopathologic findings and associated severities in the gonads of fathead minnows exposed to 17β-estradiol.
Histopathologic finding Range of severity Testis Cellular debris in efferent ducts Minimal-to-mild
Germ cell syncytia Minimal
Granulomas Minimal
Mineralization Minimal
Mineralization in efferent ducts Moderate
Sperm necrosis Minimal-to-mild
Ovary Follicular atresia Minimal-to-moderate
pearance (i.e., more similar to published photomicrographs) when compared to the glutaraldehyde/GMA sections. Forma- lin fixation was considered to be adequate for FHM ovarian follicle differentiation, although this technique was associ- ated with minor artifactual distortions such as wrinkling of the vitelline envelope and dissociation of the follicles. Early evidence of oocyte atresia, such as minor perforations of the vitelline membrane, was more readily observed in the Bouin’s-fixed sections when compared to sections fixed in formalin.
Histopathologic abnormalities (Table 3) that were diag- nosed in the testes of male FHM were generally mild and included cellular debris in spermatic ducts, granulomas, min- eralization within the germinal epithelium, mineralization in spermatic ducts, germ cell syncytia, and sperm necrosis. The sole histopathologic finding in female fish was follicular atre- sia. Atretic follicles characterized in the next section.
There appeared to be minimal, if any, damage to the testes or ovaries as a function of tissue collection. For example, although collapsed follicles were evident in all ovaries to varying degrees, such follicles almost invariably had changes within their supportive cells (e.g., intracytoplasmic yolk ma- terial, perifollicular macrophage aggregates) that suggested that their collapse was caused by antemortem degeneration (atresia) versus postmortem handling trauma.
Quantitative Histomorphologic Assessment
The goals of counting a minimum of 2,000 cells (covered by grid intersection points) for each male fish and 100 follicles for each female fish were easily achieved. The average num- ber of tagged cells per male fish (both testes combined) was 2,712, and the average for females (both ovaries combined) was 275. The relative percentages of gonadal cell types per fish as obtained by manual tagging are presented graphically (Figures 2 and 3). The ability of the pathologist to magnify each image to the limits of image resolution was beneficial for the identification of individual cell types. This degree of magnification would not have been available if counts had been performed directly from glass slides using a 40×mi- croscope objective.
As previously mentioned, manual tagging of images ac- quired from the testis (Figure 4) utilized glutaraldehyde/
GMA sections.Spermatozoawere readily identified as small (approximately 2 µm diameter), dense, discrete, approxi- mately circular, essentially acytoplasmic, deep basophilic cells that were present in large numbers within the lumina of spermatogenic lobules. Compared to spermatozoa,Sper- matidswere slightly larger (approximately 2–3µm diame- ter) than spermatozoa, and had narrow rims of eosinophilic
FIGURE1.—(A) Testis, formalin/paraffin method, H&E stain. Identification of cell types is seriously hampered by artifactual shrinkage and condensation of nuclei.
Formalin was considered to be an unsatisfactory fixative for testis cell identification. Note the presence of vacuolated cells (VC=short arrow) and apoptotic body cells (ABC=long arrow). (B) Ovary, formalin/paraffin method, H&E stain. This method is adequate for germ cell identification despite minor artifactual changes including dissociation of the follicles, contraction of the oocytes, and wrinkling of the vitelline membrane (arrow). Recognition of early atretic changes is difficult. (C) Testis, Bouin’s paraffin method, H&E stain. Nuclear detail of germinal cells is much improved compared to the formalin/paraffin method, but there is overlap of cells due to the thickness of the tissue section (3–5µm), and cytoplasmic detail is lacking (VC=short arrow; ABC=long arrow). (D) Ovary, Bouin’s/paraffin method, H&E stain. There is excellent contrast and detail, and early atretic changes can be readily identified. For the ovary, this was considered to be the most satisfactory of the evaluated preparation methods. (E) Testis, glutaraldehyde/GMA method, H&E stain. The thin sectioning (1–2µm) afforded by this method provides superior nuclear and cytoplasmic detail. This level of detail is requisite for the classification of individual spermatogenic cells in situations where the cellular morphology may be altered by exposure of the fish to test compounds (VC=short arrow; ABC=long arrow). (F) Ovary, glutaraldehyde/GMA method, H&E stain. Compared to Bouin’s/paraffin method, there is diminished color contrast among the various follicle types. Because of the relatively large size of even the smallest follicles, thin sectioning offers little advantage for ovarian cell type identification.
Figures 2–4
FIGURE2.—Relative percentages of germinal cells in the testes of male fathead minnows exposed to 17β-estradiol as obtained by manual tagging. 3.—Relative percentages of germinal cells in the ovaries of female fathead minnows exposed to 17β-estradiol as obtained by manual tagging. 4.—Testis, glutaraldehyde/GMA method, H&E stain. Manual tagging of digitized photomicrographic images is illustrated. A grid of green boxes extends across the image, and cell-type differentiation is indicated by colored dots that are individually applied to each box using the computer mouse. Boxes that remain untagged represent interstitial areas, areas devoid of tissue, or cells that cannot be reasonably classified due to suboptimal orientation or marked degeneration. Cell counts (>2,500 cells were counted per male fish) are automatically exported to a spreadsheet for tabulation (not shown).
cytoplasm. Spermatids were present in small clusters or individually within the superficial germinal epithelium; occa- sionally, spermatids were free within the lobular lumen.Sper- matocytes(primary and secondary, approximately 4–8µm diameters) were usually recognized as well-defined nests of cells within the superficial germinal epithelium. Compared to spermatids, spermatocytes had larger, somewhat less densely basophilic nuclei with a smooth chromatin pattern, inappar- ent or very faint nucleoli, and small amounts of very pale cytoplasm with polygonal shaped borders. Spermatogonia (approximately 5–10µm in diameter) were present in small clusters or individually within the superficial or deep germi- nal epithelium (Figure 5). These cells had large, pale, poorly demarcated, open-faced nuclei, nucleoli that tended to be very prominent, and abundant grainy amphophilic cytoplasm that frequently contained scattered small red perinuclear granules (this degree of cellular detail was only apparent in speci- mens that were preserved in glutaraldehyde and embedded in GMA).
Vacuolated cells (VC) were either observed within the germinal epithelium in random locations or in the lobular lumen as protrusions from the germinal epithelium. VCs were generally large cells that had variable amounts of cy- toplasmic vacuolization. At 1 end of the spectrum, some vacuolated cells had 1 to several large clear cytoplasmic vac- uoles that did not displace the nucleus (which usually resem- bled a spermatogonium or spermatocyte nucleus; Figure 5);
whereas, at the opposite end of the spectrum, other vacuo- lated cells had greatly expanded, pale, foamy cytoplasm and marginated, condensed, sharply angular, hyperchromatic nu- clei (Figure 6).Apoptotic body cells (ABC)(Figure 5) con- tained one or more variably sized spherical bodies (basophilic or basophilic with an eosinophilic rim) within one or more large cytoplasmic vacuoles.
As previously mentioned, manual tagging of images ac- quired from the ovary (Figure 7) utilized Bouin’s/paraffin sections. In such sections, perinucleolar follicles (approx- imately 50–300µm diameter) were generally the smallest follicles and were characterized by a thick zone of granular, dark basophilic ooplasm that surrounded a sharply defined core of flocculent pale lavender karyoplasm. Often apparent within the margins of the karyoplasm were minute, spherical pink bodies (provitelline nucleoli). Some of the larger per- inucleolar follicles had slightly paler ooplasm that contained a scattering of small (less than 10 µm diameter) predomi- nately clear, cytoplasmic vacuoles (developing cortical alve- oli).Cortical alveolar follicles(approximately 100–500µm diameter) were generally larger than perinucleolar follicles and were characterized by amphophilic ooplasm that con- tained numerous, predominately pale pink cytoplasmic bod- ies (cortical alveoli). The cortical alveoli stage was the first stage in which the refractile eosinophilic vitelline envelope was clearly evident. Nuclei (germinal vesicles) were usu- ally eosinophilic at this stage of development and were often crowded centrally by the cortical alveoli.Early vitellogenic follicles(approximately 250–500µm diameters) resembled cortical alveolar follicles in size and substance, except that the ooplasm of the former contained a small-to-moderate amount of bright eosinophilic, granular-to-globular material (yolk granules). Inlate vitellogenic follicles(approximately 500–750µm diameter), the eosinophilic yolk granules were
larger and more numerous compared to those of early vitel- logenic follicles, and yolk granules filled the central area of the follicle. The central yolk area of midvitellogenic folli- cles was enveloped by a zone of blue-gray ooplasm that was greater than 50µm in width. Due to an increase in yolk mate- rial, the width of this blue-gray zone was decreased or almost inapparent inmature/spawning follicles(approximately 750–
1,000 µm diameter). The most subtle (earliest) indications that follicles wereatretic (degenerative) were alterations in the morphologic appearance of the vitelline envelope that became thickened, clumped, less refractile, and developed multiple perforations (Figure 8). These changes were often accompanied by alterations in the adjacent perifollicular cells that included columnar swelling, cytoplasmic vacuolization, and the presence of intracytoplasmic droplets of eosinophilic (yolk) material. Most of the atretic follicles appeared to be in the later stages of vitellogenesis prior to their degenera- tion, although occasional cortical alveolar follicles and rare perinucleolar follicles were also atretic.
DISCUSSION
In this study, the FHM were exposed to a high concen- tration of 17β-estradiol to simulate a “worst-case scenario”
with regard to alterations in gonadal morphology. The dose of the test article (10 nM 17β-estradiol) and duration of expo- sure (14 days) were derived from previously published work (Miles-Richardson, 1999a) in which this regimen was shown to produce moderate pathologic alterations in the gonads in- cluding degenerative changes and Sertoli cell proliferation (testis) and oocyte atresia (ovary). Notably, the use of this regimen ensured that severe tissue compromise (e.g., overt necrosis) would be avoided, as severe changes might exces- sively hinder gonad cell identification and possibly affect survival of the fish.
Fathead minnows were chosen as test subjects because they are an archetypal member of the ecologically important and ubiquitous Cyprinidaefamily, and because they have been used extensively in various short-term assays, chronic life- cycle studies, and early life-stage survival and development tests involving endocrine-active substances in support of reg- ulatory programs in both North America and Europe (Ankley et al., 2001; Jensen et al., 2001; Panter et al., 2002, U.S.
EPA, 2002). Reproductive maturity in FHM occurs within 4 to 5 months of hatch under optimal conditions (U.S. EPA, 2002). FHM are fractional spawners; consequently, gameto- genesis in both sexes occurs asynchronously, which signifies that the various commonly recognized gametogenic cell types are usually evident simultaneously (U.S. EPA, 2002).
As in most teleosts, the structural organization of the FHM testis is of the unrestricted type, in which spermatogene- sis occurs within the germinal epithelium along the entire length of the seminiferous lobule (Grier, 1976; Grier, 1981;
Pudney, 1993; U.S. EPA, 2002). The functional unit of the testicular lobule is the spermatocyst, wherein a clonal ag- gregate of precursor cells in a synchronous phase of devel- opment is completely surrounded by the cytoplasmic arms of a single Sertoli cell (Grier, 1993; Pudney, 1993; Pati˜no and Sullivan, 2002). Following establishment of an associ- ation between a spermatogonium and a Sertoli cell, sper- matogenesis progresses through a series of mitotic and mei- otic divisions that sequentially produce smaller cells and
Figures 5–7
FIGURE5.—Testis, glutaraldehyde/GMA method, H&E stain. Spermatogonia (SG) have large ovoid nuclei, peripherally clumped chromatin, prominent nucleoli, and moderate-to-abundant amounts of granular cytoplasm. It is presumed that most of these cells are primary and secondary spermatogonia, but some may be hypertrophic Sertoli cells. The long arrow indicates an apoptotic body cell that contains five spherical structures within a large cytoplasmic vacuole. Two other vacuolated cells (VC) in this image (short arrows) are not overtly phagocytic, and some of these cells may be degenerating spermatogonia or spermatocytes. Note that even spermatozoa (SZ) are viewed as a monolayer in 1–2µm sections. 6.—Testis, glutaraldehyde/GMA method, H&E stain. The large arrows indicate vacuolated cells (VC) with voluminous foamy cytoplasm and relatively small, angular, condensed nuclei. Some of these cells appear to contain spermatozoa (small arrows) within their cytoplasm. SZ=spermatozoa, SC=spermatocytes, SG=spermatogonia. 7.—Ovary, Bouin’s/paraffin method, H&E stain. Manual tagging of digitized photomicrographic images is illustrated. Due to the small number of cells in the ovary relative to the testis, a grid overlay is not utilized, and all germinal cells are counted. A different colored square represents a single stage of follicular development.
FIGURE8.—Ovary, Bouin’s/paraffin method, H&E stain. Shown are a late vitellogenic follicle (LV), an early atretic follicle (EA), and a late atretic follicle (LA) follicle. In the early atretic follicle, the vitelline envelope (ve) is clumped and irregular, and small pores are evident (long arrow). Such changes are exaggerated in the late atretic follicle. Note that the perifollicular cells (arrowheads) of the mid-vitellogenic follicle are small and flattened; whereas, the perifollicular cells (short arrows) of the early atretic follicle are columnar, vacuolated, and have intracytoplasmic droplets.
a haploid chromosome configuration according to the fol- lowing sequence: spermatogonia→spermatocytes→sper- matids (Grier, 1981; Nagahama, 1983; Pudney, 1995; U.S.
EPA, 2002). Spermatogonia and spermatocytes are often sub- divided into primary and secondary developmental stages (Grier, 1981; Nagahama, 1983; Pudney, 1995). Spermatids undergo maturation events (spermiogenesis) that include nu- clear shape changes, development of a flagellum, loss of cyto- plasmic volume, and phagocytosis of cytoplasmic remnants (residual bodies) prior to their release into the lobular lu- men as spermatozoa (spermiation) (Pudney, 1995; Pati˜no and Redding, 2000).
In the piscine ovary, the functional unit is the follicle that consists of an oocyte surrounded by perifollicular cells (gran- ulosa cells and cells of the thecal layer) (Nagahama, 1983;
Tyler and Sumpter, 1996; Pati˜no and Redding, 2000). Mi- totic and meiotic divisions during oogenesis halve the chro- mosome number and generate progressively larger cells in the following manner: oogonia → perinucleolar (primary growth) oocytes→cortical alveolar oocytes→early vitel-
logenic oocytes → late vitellogenic oocytes → mature/
spawning oocytes (U.S. EPA, 2002). Some authors have fur- ther subdivided early primary growth follicles into additional developmental stages (e.g., bouquet, chromatin-nucleolus, etc.) (Pati˜no and Redding, 2000). Vitellogenesis, the pro- cess by which extra-ovarian proteins produced by the liver are packaged into oocytes, is a hormonally controlled activity that results in tremendous follicular enlargement (Tyler and Sumpter, 1996; Pati˜no and Sullivan, 2002). Following the maturation period, ovulation occurs when oocytes are shed from within their perifollicular cell sheath and are expelled into the ovarian lumen in arrested phase of the second mei- otic division (Pati˜no and Sullivan, 2002). Follicular atre- sia, whether caused by physiologic stimuli or environmen- tal stress, entails the degeneration of follicles at any stage of follicular development (Nagahama, 1983) although atretic follicles are often most noticeable in the later developmen- tal stages (Tyler and Sumpter, 1996). Follicular atresia is accompanied by hypertrophy of the perifollicular granulosa cells (Nagahama, 1983).
Gonad Excision
The microdissection approach used to excise the FHM male and female gonads was a rapid and straightforward tech- nique that was easily standardized. Despite the diminished size and/or altered texture of some of the gonads from these estradiol-treated FHM, this method consistently yielded in- tact, well-preserved reproductive organs that were essentially devoid of collection-induced microscopic artifacts. Prior to this experiment, longitudinal (sagittal and parasagittal) or transverse whole fish sectioning was briefly considered as an alternative technique to gonad excision. Although whole fish sectioning might be appropriate for gonad cell quantifi- cation in smaller laboratory fishes such as Japanese medaka Oryzias latipesand zebrafishDanio rerio, there are several reasons why such methods are less ideal for larger fishes such as FHM. First, even if the head and the tail were to be removed, standard plastic tissue cassettes are not spacious enough to accommodate the trunk segments of some large adult FHM males if positioned longitudinally.
Although oversize cassettes are available, microtoming the larger paraffin blocks is technically more difficult compared to standard block sizes. Second, because FHM gonads are elongated tubular structures that are oriented essentially par- allel to the longitudinal axis of the fish, and since perfectly level embedding of the whole body is difficult to achieve, the entire length of a gonad cannot be consistently recovered on a single slide using longitudinal whole body sectioning. Con- sequently, a great deal of animal-to-animal variability in both the amount of tissue recovered and the plane of sectioning are to be expected.
Another drawback is that it may be difficult to accurately identify individual gonads as right or left when parasagit- tal (vertical) whole body sectioning is performed (i.e., if the plane of microtoming is slightly oblique, segments of both gonads may appear in the same tissue section), and at least 2 whole body sections must be prepared per fish if both the right and left gonads are to be evaluated. This draw- back would be alleviated if whole fish were sectioned in the frontal (horizontal) plane; however, because FHM normally have an anatomic profile that is laterally compressed, and the dorsal and ventral surfaces of these fish are not parallel to their longitudinal axis, achieving a consistent plane of section would be extremely challenging (if not impossible) using this technique.
On the other hand, if whole fish were microtomed trans- versely instead of longitudinally, numerous sections would need to be produced and evaluated in order to ensure ad- equate spatial representation of the gonads. A third reason to prefer gonad excision is that this technique avoids mi- crotoming artifacts that can be caused by the presence of firmer tissues (such as bone and cartilage) that are inherently present in whole body sections. Finally, gonad excision al- lows these organs to be visualized macroscopically in situ, and it is possible to obtain gonad weights if desired. Con- cerning the cost-effectiveness of this technique, it should be noted that amount of time required to excise the gonads is only slightly greater than the time that would take to care- fully incise the abdomen and flush it with fixative; in addition, decalcification procedures can be omitted if the gonads are excised.
Fixation and Embedding Procedures
For quantitative gonad staging, the importance of appro- priate microdissection procedures and an optimal combina- tion of histological procedures cannot be overemphasized.
Compared to semiquantitative grading, in which the relative density of cell populations is subjectively estimated, a higher degree of precision is required to be able to identify and quan- tify gametogenic precursors on a cell-by-cell basis. This is a prerequisite for the testis, as cellular detail is intrinsically lim- ited by the diminutive size of the germinal cells (the largest spermatogenic cells are approximately 10µm diameter). For the FHM testes in this study, the thin sectioning (2µm or less) that is possible with glutaraldehyde-fixed/GMA-embedded sections greatly improved visualization of nuclear and cyto- plasmic details in what was essentially a monolayer of cells.
Other reported benefits of glutaraldehyde fixation and GMA embedding for preparing FHM gonad sections include im- proved tissue fixation, reduced shrinkage of sections, reduced distortion due to microtoming, and ease of staining (U.S.
EPA, 2002). Because ovarian follicles are vastly larger than spermatogenic cells, thin sectioning was not particularly ad- vantageous for the ovary. Nevertheless, the visualization of early atretic changes in the follicles was enhanced by Bouin’s fixation when compared to formalin fixation, which again il- lustrates that the correct choice of histological procedures is crucial for the accurate assessment of microscopic changes.
The various types of macroscopic and histopathologic find- ings observed in the male and female FHM gonads in this study (Tables 2 and 3) were not considered to be unique to 17β-estradiol-treated fish; however, it was likely that overall severity and/or prevalence of some of these findings was a consequence of 17β-estradiol administration, as suggested earlier (Miles-Richardson et al., 1999a). For example, while follicular atresia is considered to be an integral feature of the normal teleost ovarian cycle (Pati˜no and Redding, 2000) the proportion of atretic follicles identified in these FHM (mean=18.9%, range=7.4–26.9%, n=3) was higher than anticipated. Since reported mean percentages of atretic folli- cles in groups of control FHM have varied from 1.6% (Mc- Cormick et al., 1989) to 5% (Miles-Richardson et al., 1999b), and it has been speculated that follicular atresia up to 10–12%
may be normal for some FHM (U.S. EPA, 2002). Similarly, vacuolated cells (VC) and apoptotic body cells (ABC) are also frequently present in the testes of FHM that have not been treated with estrogenic compounds (personal observa- tion), although the numbers of such cells in the current study were generally higher than what we would expect to see in untreated fish.
Based on prior experience, the presence of occasional granulomas and mineralized foci in the testes of the male FHM was not surprising; whereas, the minimal to mild sperm necrosis (observed unilaterally or bilaterally in all 3 fish) was considered to be a more unusual finding. Sperm necrosis was also reported to occur in male FHM that were exposed to 4- nonylphenol and nonylphenol ethoxylate (Miles-Richardson et al., 1999b). The testis attenuation that was observed at necropsy in some of the male FHM could not be correlated with specific histopathologic lesions; most likely, the atten- uation was due to an overall decrease in the cellularity (cell density) of the testes in these particular fish.
One gonadal alteration that has been reported in other studies is the presence (or relatively increased incidence) of oocytes within the testes of male teleosts that have been ex- posed to estrogenic substances (Gray and Metcalfe, 1997;
Metcalfe et al., 2001). In 1 study, in which a lake was experi- mentally treated with the synthetic estrogen, ethynylestradiol, the incidence of ovotestis formation in resident male FHM was 11%. (Palace et al., 2002). Oocytes were not observed in the testes of male FHM in the present study, possibly due to the short duration of the exposure (10 days), the exposure of sexually mature fish, and/or the limited number of fish that were examined. For future studies, the manual tagging pro- cedure could be easily adapted to count oocytes that might be present within testis sections.
Overview of FHM Testis and Ovary Cell Types
Prior to manually tagging the gonadal cells in digital im- ages, it was necessary to define the morphologic criteria for cell categorization. Although various descriptions of FHM gametogenic cells have been published, (Miles-Richardson et al., 1999a), none were considered to be completely satis- factory for counting individual cells at the light microscopic level. Regarding the testis, cell categorization was more chal- lenging than anticipated, as there appeared to be disparity in the scientific literature concerning classification criteria.
Particularly problematic was the issue of Sertoli cell histo- morphology. While there is considerable information pertain- ing to Sertoli cell-germ cell interactions in mammalian testis (Russell and Griswold, 1993), and while the fundamental roles of mammalian and fish Sertoli cells appear to be com- parable (Russell and Griswold, 1993), only limited data were available for fish in general (Grier, 1993); and fish demon- strated far more functional and morphological diversity in their reproductive strategies when compared to mammals (Grier, 1993). Of special importance in the present context was the fact that abnormally large Sertoli cells (hypertro- phy), often in parallel with increased numbers (hyperplasia), have been associated with exposure to some estrogenically active compounds (Miles-Richardson et al., 1999a, 1999b;
Kinnberg et al., 2000; van der Ven et al., 2002; Kinnberg and Toft, 2003).
Although there is a consensus description among vari- ous sources in which the normal Sertoli cells of various osteichthyes are characterized by small pyramidal-shaped or elongated nuclei and minimal or attenuated cytoplasm (Pudney, 1993; Miles-Richardson et al., 1999a; Kinnberg et al., 2000; U.S. EPA, 2002), the light microscopic ap- pearance of altered Sertoli cells as represented in the sci- entific literature is ambiguous. For example, in 1 report in which Sertoli cell hypertrophy in FHM was considered to be a consequence of 17β-estradiol exposure, the hypertrophic Sertoli cells were depicted as having a “vesiculate nucleus and a prominent nucleolus” (Miles-Richardson et al., 1999a).
Based on this description and associated photographs, the enlarged Sertoli cells in this paper had a polygonal shape and appeared virtually indistinguishable from spermatogo- nia. Pudney, for example, characterized spermatogonia as having “large regular nuclei usually containing a distinct nu- cleolus” (Pudney, 1995).
Conversely, in the testes of zebrafishDanio rerioexposed to 17α-methyldehydrotestosterone (van der Ven et al., 2003),
the hypertrophic Sertoli cells appeared to be proportionally smaller cells that had ovoid to reniform nuclei, minimal ob- vious cytoplasm, and variably evident nucleoli. In another study, “hypertrophic cuboidal to columnar Sertoli cells with spherical nuclei” were observed in the testes of control platy- fishXiphophorus maculatus, and increased numbers of these cells were present in the testes of platyfish that had been ex- posed to nonylphenol (Kinnberg et al., 2000). Concerning the physiologic hypertrophy of Sertoli cells as is reported to occur in poecilids, these cells have also been described as columnar (Pudney, 1995). In the photomicrographs that accompanied these reports, the hypertrophic Sertoli cells of zebrafish and platyfish bore little resemblance to spermatogonia.
A measure of inconsistency in the characterization of hy- pertrophic Sertoli cells is undoubtedly due to interspecies dif- ferences. It is also possible that the hypertrophic appearance of some Sertoli cells may be dependent on the nature of the cytomegalic stimulus (i.e., the morphology of these cells may vary according to the physiological requirements of the testis during a given period of development or according to the type of exogenous chemical exposure). An additional considera- tion is the fact that piscine Sertoli cells have few defining light microscopic features. Even at the ultrastructural level, justification for the homologous existence of Sertoli cells in fishes is primarily dependent upon the finding of residual spermatogenic materials in the cytoplasm of such cells and the presence of junctional complexes that define the Sertoli cell barrier (Grier, 1981, 1993). Thus, it is not surprising that fish Sertoli cells have been confused with other cell types, such as spermatogonia, when viewed by light microscopy (Pudney, 1993).
Currently, ultrastructural analysis appears to be the method of choice for the definitive identification of piscine Sertoli cells. Unfortunately, electron microscopy is an impractical tool for assessing cell proportions because the number of cells that can be viewed at any one time is very limited. Until further standards of recognition become available (e.g., the identification of Sertoli cells via routine immunohistochemi- cal staining), the most rational approach may be to utilize de- scriptive terminology for those cells that cannot be associated with well-defined cell categories by light microscopy. On this basis, and for the purpose of cell counting in this study, the label “spermatogonium” was applied to cells that possessed particular morphologic features, including large overall size, large, poorly demarcated, open-faced nuclei, prominent nu- cleoli, eosinophilic perinuclear granules, and moderate-to- large amounts of grainy cytoplasm (Figures 4 and 5). It is likely that the vast majority of these cells were spermato- gonia (primary and secondary); however, based on various descriptions and photographs in the scientific literature, it is also possible that at least some of these were hypertrophic Sertoli cells. For each of the 3 separate 17β-estradiol-treated male fish in this study, a potentially important result from the manual tagging exercise was the relative predominance of spermatogonia when compared to spermatocytes and sper- matids (Figure 2). It should also be noted that relatively low numbers of small, angular cells consistent with nonhyper- trophic Sertoli cells were recognized during the manual tag- ging procedure but were not specifically tagged.
In addition to ambiguity regarding the morphology of hy- pertrophic Sertoli cells, the categorization of testis cell types
was challenging because previously established staging sys- tems did not specifically address other forms of histomorpho- logically altered cells. In particular, 2 types of altered cells that were observed in this study were termed vacuolated cells (VC) and apoptotic body cells (ABC). The cellular origins of VC and ABC were not always obvious during light mi- croscopic examination; consequently, to justly labeling such cells, descriptive appellations such as VC and ABC were cre- ated for cell counting purposes. The general appearance of many VC (those with 1 to several large, clear, cytoplasmic vacuoles that did not displace the nucleus) strongly suggested that they were vacuolated spermatogonia or spermatocytes that had undergone degenerative changes as a consequence of 17β-estradiol exposure (Figure 4). It is quite possible that other VC (those with greatly expanded, pale, foamy cyto- plasm) were derived from a different cell type (e.g., activated macrophages or Sertoli cells) (Figure 5).
The presence of large vacuoles in Sertoli cells has been at- tributed to phagocytosis of residual bodies and other cellular elements (Loir et al., 1995). Although it could be contended that VC were artifacts of tissue handling or processing in this experiment, this notion is difficult to defend as VC (and ABC) were evident in all 3 fixation/embedding combinations and in each testis to varying degrees (Figure 1). The existence of apoptotic bodies within the cytoplasmic vacuoles of some cells (ABC) supports the concept that phagocytic cells are present (and probably resident) within the testicular lobule.
It is relevant that among Chordata, a unifying function of Ser- toli cells, is the phagocytosis of effete spermatozoa and sper- matogenesis by-products (Grier, 1993; Russell and Griswold, 1993; Loir et al., 1995), and that degenerate and presump- tively apoptotic sperm have been observed ultrastructurally within the Sertoli cells of FHM and other fishes (Grier, 1993;
Miles-Richardson et al., 1999a, 1999b).
Manual Tagging of Individual Cell Types
Automated image analysis techniques, in which the iden- tification of individual cell types is performed by computer software using colorimetric and morphometric thresholds, were given initial consideration for this study. It is reason- able to assume that the throughput advantage of automated counting would be significant as compared to manual count- ing, and that it might be possible to enhance cell type dis- crimination for automated counting through the use of im- munohistochemical procedures. Ultimately, automated anal- ysis techniques were not employed, however, because cer- tain gonadal cell types (spermatogonia, VC, ABC, and many atretic follicles) would have been difficult for the image anal- ysis software to recognize as distinct entities, due to the variability and subtlety of the structural changes in these cells.
A different manual approach to quantitative testis cell eval- uation was recently reported by van der Ven, Wester, and Vos (2003). In their study, which utilized zebrafishDanio re- rio, clonal aggregates of spermatogenic cells (spermatocysts) were outlined in digitized images, and the 2-dimensional ar- eas of the outlined regions were then measured. Although this appeared to be an elegant approach for the zebrafish testes, it would have been difficult to apply this particular technique in the present study, because the organizational pattern of the FHM spermatocysts appeared to be disrupted
(i.e., spermatocysts were generally small and very poorly defined), presumably as a result of 17β-estradiol treatment.
Additionally, cell types that were present individually or in small aggregates in the FHM testes (such as ABC and VC) would not have been recognized and counted using the van der Ven, Wester, and Vos method. On the other hand, due to mi- croanatomical interspecies differences, the manual tagging protocol employed in the present study could not be read- ily utilized for certain other fishes without modification. An example would be atheriniform fishes such as the Japanese medakaOryzias latipes. Unlike FHM, in which the cell type distribution tends to be relatively homogenous throughout the length and width of the testis, spermatogenic primordia are re- stricted to the periphery of the testis in medaka (Grier, 1976).
Because the distribution of spermatogenic cells is nonuni- form in medaka, it is unlikely that a single longitudinal tissue section would adequately represent the cellular composition of an entire medaka testis.
As performed in the present study, the manual tagging of individual cell types in digital images afforded a rigorous technique for the evaluation of potential changes in gonad cell proportions, as have been reported to occur following expo- sure of fish to certain xenobiotic chemicals. Contrasted to a previous study in which FHM testis cells were counted during direct microscopic visualization of tissue sections (Sohoni et al., 2001), the image-based manual tagging technique pro- vided a permanent and reviewable record of each cell counted.
Additionally, the manual tagging technique provided bene- fits that may not be obvious at first glance. For example, the grid markings that were applied to each glass slide, and the virtual grid that was applied to each image, effectively pre- vented duplicate evaluation of testis areas and duplicate cell counting, respectively. The ability to magnify each image to the limits of image resolution was conducive to the detailed morphological inspection of the gonadal cell types. Repre- sentative sampling of the gonads was achieved by counting large numbers of cells for each fish (>2,500 cells per male and>250 cells per female).
The classification systems that were employed to identify and quantify gonadal cells utilized a flexible combination of established and descriptive terminology; thus, cells that may have been altered by 17β-estradiol exposure and there- fore were difficult to associate with defined cell categories of testicular development were readily incorporated as descrip- tive cell categories (VC, ABC). Finally, the light microscopic evaluation of individual gametogenic cells was greatly facil- itated by specimen preparation techniques that included the excision of gonads via microdissection and by optimized fix- ation and embedding procedures (5% glutaraldehyde/GMA for the testis and Bouin’s solution/paraffin for the ovary). Re- sults of this pilot study indicate that quantitative staging by manual tagging is a feasible and meticulous scientific tool for the comparative histological assessment of fathead minnow gonads.
REFERENCES
Ankley, G. T., Jensen, K. M., Kahl, M. D., Korte, J. J., and Makynen, E. A.
(2001). Description and evaluation of a short-term reproduction test with the fathead minnow (Pimephales promelas).Environ Toxicol and Chemi 20(6), 1276–90.
Ankley, G. T., Kahl, M. D., Jensen, K. M., Hornung, M. W., Korte, J. J., Makynen, E. A., and Leino, R. L. (2002). Evaluation of the aromatase inhibitor fadrozole in a short-term reproduction assay with the fathead minnow (Pimephales promelas).Toxicol Sci67, 121–30.
Christiansen, T., Korsgaard, B., and Jespersen,A. (1998). Effects of nonylphe-a nol and 17β-estradiol on vitellogenin synthesis, testicular structure and cytology in male eelpout (Zoarces viviparus).J Exp Biol 201, 179–
92.
Department of the Interior, U.S. Geological Survey, Biological Resources Di- vision. (2000).U.S. Biomonitoring of Environmental Status and Trends (BEST) Program: Selected Methods for Monitoring Chemical Contami- nants and Their Effects in Aquatic Ecosystems. Information and Technol- ogy Report, pp. 30–42 USGS/BRD/ITR—2999-0005. Washington, DC.
Gray, M. A., and Metcalfe, C. D. (1997). Induction of testis-ova in Japanese medaka (Oryzias latipes) Exposed top-Nonylphenol.Environ Toxicol and Chem16, 1082–6.
Grier, H. J. (1976). Sperm development in the teleostOryzias latipes.Cell Tiss Res168, 419–31.
Grier, H. J. (1981). Cellular organization of the testis and spermatogenesis in fishes.Am Zool21, 345–57.
Grier, H. J. (1993). Comparative organization of Sertoli cells including the Sertoli cell barrier. InThe Sertoli Cell(L. D. Russell and M. D. Griswold, eds.), pp. 704–739. Cache River Press, Clearwater, FL.
Jensen, K. M., Korte, J. J., Kahl, M. D., Pasha, M. S., and Ankley, G. T. (2001).
Aspects of basic reproductive biology and endocrinology in the fathead minnow (Pimephales promelas).Comp Biochem Physiol C Toxicol Phar- macol.128, 127–41.
Kinnberg, K., Korsgaard, B., Bjerregaard, P., and Jespersen,A. (2000). Effectsa of nonylphenol and 17β-estradiol on vitellogenin synthesis and testis mor- phology in male platyfish (Xiphophorus maculatus).J Exp Biol203, 171–
81.
Kinnberg, K., and Toft, G. (2003). Effects of estrogenic and antiandrogenic compounds on the testis structure of the adult guppy (Poecilia reticulata).
Ecotoxicol Environ Saf54, 16–24.
Loir, M., Sourdaine, P., Mendis-Handagama, S. M. L. C., and Inra, B. J. (1995).
Cell-cell interactions in the testis of teleosts and elasmobranchs.Microsc Res Tech5(32), 533–52.
McCormick, J. M., Stokes, G. N., and Hermanutz, R. O. (1989). Oocyte atresia and reproductive success in fathead minnows (Pimephales promelas) ex- posed to acidified hardwater environments.Arch Environ Contam Toxicol 18, 207–14.
Metcalfe, C. J., Croley, T. R., March, R. E., and Potter, T. (2001). Estrogenic potency of chemicals detected in sewage treatment plant effluents as deter- mined by in vivo assays with Japanese medaka (Oryzias latipes).Environ Toxicol Chem20, 297–308.
Miles-Richardson, S. R., Kramer, V. J., Fitzgerald, S. D., Render, J. A., Yamini, B., Barbee, S. J., and Giesy, J. P. (1999a). Effects of waterborne exposure of 17β-estradiol on secondary sex characteristics and gonads of fathead minnows (Pimephales promelas).Aquat Toxicol47, 129–45.
Miles-Richardson, S. R., Pierens, S., Nichols, K. M., Kramer, V. J., Snyder, E. M., Snyder, S. A., Render, J. A., Fitzgerald, S. D., and Giesy, J. P. (1999b). Effects of waterborne exposure to 4-nonylphenol and nonylphenol ethoxylate on secondary sex characteristics and gonads of fathead minnows (Pimephales promelas). Environ Res 80, S122–
S137.
Nagahama, Y. (1983). The functional morphology of teleost gonads. InFish Physiology(W. S. Hoar, D. J. Randall, and E. M. Donaldson, eds.), pp. 223–
275. Academic Press, San Diego, CA.
Nichols, K. M., Miles-Richardson, S. R., Snyder, E. M., and Giesy, J. P. (2001).
Effects of exposure to municipal wastewater in situ on the reproductive physiology of the fathead minnow (Pimephales promelas).Environ Toxicol and Chem18, 2001–12.
Palace, V. P., Evans, R. E., Wautier, K., Baron, C., Vandenbyllardt, L., Vander- steen, W., and Kidd, K. (2002). Induction of vitellogenin and histological effects in wild fathead minnows from a lake experimentally treated with the synthetic estrogen, ethynylestradiol.Water Qual Res J Canada37, 637–50.
Panter, G. H., Hutchinson, T. H., L¨ange, R., Lye, C. M., Sumpter, J. P., Zerulla, M., and Tyler, C. R. (2002). Utility of a juvenile fathead minnow screening assay for detecting (anti-) estrogenic substances.Environ Toxicol Chem21, 319–26.
Pati˜no, R., and Redding, J. M. (2000). Reproductive systems. InThe Laboratory Fish(G. K. Ostrander, ed.), pp. 489–500. Academic Press, San Diego, CA.
Pati˜no, R., and Sullivan, C. V. (2002). Ovarian follicle growth, maturation, and ovulation in teleost fish.Fish Physiol Biochem26, 57–70.
Pudney J. (1993). Comparative cytology of the non-mammalian vertebrate Ser- toli cell. InThe Sertoli Cell(L. D. Russell and M. D. Griswold, eds.), pp. 612–657. Cache River Press, Clearwater, FL.
Pudney, J. (1995). Spermatogenesis in nonmammalian vertebrates.Microsc Res Tech32, 459–97.
Russell, L. D., and Griswold, M. D. (1993).The Sertoli Cell. Cache River Press, Clearwater, FL.
Smith, R. J. F. (1978). Seasonal changes in the histology of the gonads and dorsal skin of the fathead minnow,Pimephales promelas.Can J Zool56, 2103–9.
Sohoni, P., Tyler, C. R., Hurd, K., Caunter, J., Hetheridge, M., Williams, T., Woods, C., Evans, M., Toy, R., Gargas, M., and Sumpter, J. P. (2001).
Reproductive effects of long-term exposure to bisphenol A in the fathead minnow (Pimephales promelas).Environ Sci Tech35, 2917–25.
Tyler, C. R., and Sumpter, J. P. (1996). Oocyte growth and development in teleosts.Rev Fish Biol Fisheries6, 287–318.
U.S. Environmental Protection Agency. (2002). A Short-term Test Method for Assessing the Reproductive Toxicity of Endocrine-Disrupting Chemicals Using the Fathead Minnow (Pimephales promelas). EPA 600/R-01-067, pp. 141, 144. Washington, DC.
Van den Belt, K. P., Wester, W., van der Ven, L. T. M., Verheyen, R., and Witters, H. (2002). Effects of ethynylestradiol on the reproductive physiology in zebrafish (Danio rerio): time dependency and reversibility.Environ Toxicol and Chem21, 767–75.
van der Ven, L., and Wester, P. (2002). Normal histology and effects of hormonal active agents of reproductive organs of the zebrafish (Danio rerio). OECD Technical Workshop on Gonadal Histology of Small Laboratory Fishes as a Tool for Endocrine Disruption Screening and Testing. OECD Bilthoven, Netherlands.
van der Ven, L. T. M., Wester, P. W., and Vos, J. G. (2003). Histopathology as a tool for the evaluation of endocrine disruption in Zebrafish (Danio rerio).
Environ Toxicol Chem22, 908–13.
Yonkos, L. T., Kane, A. S., and Reimschuessel, R. (2000). Fathead minnow histology atlas: worldwide web outreach and utilization.Mar Environ Res 50, 1–5.
